IAPRS Volume XXXVI, Part 5, Dresden 25-27 September 2006
EXPERIMENTS ON DEFORMATION MEASUREMENTS OF“HELSI
NKIDESI
GN
WEEK2005”I
NFO PAVI
LION
Petri Rönnholma, Petteri Pöntinena, Milka Nuikkaa, Antti Suominena, Hannu Hyyppäa,
Harri Kaartinenb, Ilmari Absetza, Hannu Hirsia, and Juha Hyyppäb
a
b
Helsinki University of Technology, P.O. Box 1200, FI-02015 TKK, Finland –first.lastname@tkk.fi
Finnish Geodetic Institute, Geodeetinrinne 2, P.O. Box 15, FI-02431 Masala, Finland –first.lastname@fgi.fi
Commission V
KEY WORDS: terrestrial laser scanning, close-range photogrammetry, construction quality, deformation, CAD/CAM, load test
ABSTRACT:
The”
He
l
s
i
nk
iDe
s
i
g
nWe
e
k
”i
nf
opa
v
i
l
i
oni
sa
ne
x
a
mpl
eofachallenging mass customisation based industrial construction planned
with CAD techniques. The final pavilion was modelled using both close-range photogrammetric methods and terrestrial laser
scanning. The as-built model was compared to the design model revealing the maximum difference of 0.035 m. Two load tests (20
and 100 kg) were arranged to simulate low and moderate snow loads on a roof structure and to arrange two different loads versus
deformation cases for measurements. Both modelling methods yielded similar result indicating that the 20 kg load did not cause
significant deformations to the pavilion. In the case of 100 kg load, the rim arcs kept their shapes well, but some deformations was
found from the shell structures. The results from photogrammetry and laser scanning were compared and both methods had some
advantages and disadvantages. The strength of photogrammetry was more accurate alignment when corners were measured and the
strength of the laser scanning was the more complete surface model. In our case, the photogrammetric models were measured from
free-network image blocks. One model appeared to have some distortion along the z-axes. The disadvantages with laser data
included difficulties to measure accurate corner points and the strong deformation of reference spheres in moist conditions.
1. INTRODUCTION
As a natural hygroscopic material, wood is apt to deformations
in shape and dimensions due to changes in humidity. Also
exterior loads such as snow, wind and service loads cause
deformations in building structures. Wood is a common
material for load carrying building frame-structures as well as in
interior and exterior surface-structures. Glass is also used
frequently, especially in facades. The HDW pavilion has these
two materials in a relatively complex and challenging form.
The“
He
l
s
i
nk
iDe
s
i
g
nWe
e
k2005”(HDW) info pavilion is an
example of a complex structure using the advantages of glass
and plywood (Figure 1). The glass surface consists of 135 glass
triangles enchased with plywood ribs. The glass functions as a
stressed skin; the triangles of plywood prevent the glass from
bending and the free-shaped rim arcs supply the load to the base
(Lehto and Seppänen, 2005).
Another point of our interests of this portable structure was,
how rigid the shape of the HDW is and how to detect the
deformations when different loads were attached to the pavilion.
For this task two measuring methods were selected: close-range
photogrammetry and terrestrial laser scanning.
Close-range photogrammetry is a well-accepted tool for 3D
measurements and commonly used in various tasks such as for
cultural heritage documentation (Gruen et al., 2002), building
restoration (Cardenal et al., 2005), deformation monitoring
(Hampel and Maas, 2003; Gordon et al., 2004), and
reconstruction tasks (Grussenmeyer and Yasmine, 2003; Harun
and Ahmad, 2003; Schindler et al., 2003; Valiev, 1999). In the
most advanced applications, the 3D-model can be created from
an image sequence with a minor human interaction (Werner et
al., 2001; Pollefeys et al., 2000). Currently, it is easy to get offthe-shelf digital cameras, calibrate them and use existing
photogrammetric software for actual measurements.
Fi
g
ur
e1.“
He
l
s
i
nk
iDe
s
i
g
nWe
e
k2005”i
nf
opa
v
i
l
i
on.
The HDW Info Pavilion is a miniature of a challenging mass
customisation based industrial construction. Every single glass
triangle and plywood rib of the pavilion is unique. All of the
parts have been modelled three dimensionally in the CAD
model of the structure. Each part has been manufactured with an
automated CNC-machining (computer numerical control)
device controlled by the data from the CAD model and the
whole design-manufacturing-process was highly automated.
One point of interest was acquiring as-built information from
the pavilion and to compare the realized structure to the design
model.
In addition to the photogrammetric methods, terrestrial laser
scanning has become an established tool for acquiring 3D
models. Reported examples of applications include cultural
heritage recording (Alshawabkeh and Haala, 2004),
273
ISPRS Commission V Symposium 'Image Engineering and Vision Metrology'
architectural modelling (El-Hakim et al., 2005; Schuhmacher
and Böhm, 2005), building reconstruction (Lee and Choi,
2004), as-built documentation (Sternberg et al., 2004), forest
modelling (Thies and Spiecker, 2004), and structural
engineering (Guarnieri et al., 2005; Lindenbergh et al., 2005).
With terrestrial laser scanners, the accuracy of individual
sample points (eg. ±2 - ±5 mm) are lower than photogrammetrically acquired ones, but the model of a surface is
much more accurate because of the dense point cloud (Gordon
et al., 2004). The potential of combining close range
photogrammetry and terrestrial laser scanning is to optimise the
geometric accuracy and the visual quality of 3D data capture of
scenes (Alshawabkeh and Haala, 2002).
The photogrammetric blocks were free-networks, leaving the
possibility of scale deformations. In addition, the measured
targets were natural such as corners. The interpretation
problems of natural targets cause inaccuracies. Also the image
block geometry was not optimal, because all images were taken
close to same height level. However, the accuracy indicators
after block adjustment were not alarming.
The borders of four triangles in the corners (Figure 2) from
inside of the HDW pavilion were measured with a measuring
tape in order to get an approximate scale for the
photogrammetric model.
The main objectives of this research included comparison of
CAD plans with as-built models, measuring the temporal
deformations of wooden structures of the HDW pavilion as a
function of various sizes of loads and comparison of results
obtained by terrestrial laser scanning with those obtained by
close-range photogrammetric measurements.
The long-term scenario of this project is to develop and adapt
modern ICT based 3D measurement techniques to model and
verify the quality of construction projects and also measure the
deformations in building structures.
Figure 2. The dimensions of four corner triangles of the
pavilion were measured in order to get approximate scale for the
photogrammetric model.
2. MATERIAL
3. METHODS
During the data acquisition, the HDW pavilion was situated in
Lahti city. The main focus of observations was on interior parts,
because the outer cover of the HDW pavilion was completely
made from glass. Operating through the glass was expected to
cause too much uncertainty for distance measurements and
therefore outer cover was not included.
The initial state of the HDW pavilion was captured by the
photogrammetric image block and by making two laser scans.
The structure was loaded with a sand-filled sack lifted on the
top of the pavilion (Figure 3, left). For another test a load of
weights was hanged to the rim arc (Figure 3, right). The weights
were 20 kg and 100 kg, respectively. After adding the loads,
similar data acquisition as for initial measurements was carried
out in order to get a comparable set of data.
The applied terrestrial laser scanner was FARO LS 880 HE80,
which is based on phase measurements providing high-speed
data acquisition. Technical parameters of the scanner include
maximum measurement rate of 120000 pulses/s, wavelength of
785 nm, vertical field of view 320° and horizontal field of view
360°, and linearity error of 0.003 m (at 25 m and 84 %
reflectivity), see also www.faro.com. Accuracies and
performances of various types of laser scanners have been
reported earlier e.g. by Fröhlich and Mettenleiter (2004). Laser
scanning was done from two positions using the half resolution
for each three cases. The laser scanning data was post-processed
with the Faro Scene 3.0 and the Geomagic Studio 8. The
comparison between the CAD model and the photogrammetric
model was done in the AutoCAD.
Figure 3. Left: The HDW pavilion was stressed by lifting a 20
kg sand-filled sack on the top of the structure. Right: Another
load test included 100 kg of weights hanging from one rim of
the pavilion.
During the laserscanning, the weather conditions were not ideal.
The temperature was only a few degrees above zero (Celsius)
and the air humidity was high. During some scans, it was
slightly raining. Similar weather conditions can be expected to
be reality, when working on construction sites. Therefore, this
example was a good experiment revealing some unexpected
behaviour of the FARO laser scanner. These experiences are
described more in detail in chapter 3.
Image observations were added to the least squares block
adjustment for solving image orientations. The adjustment was
calculated using the iWitness software. An approximate scale of
the model was defined using distances from measuring tape
measurements. Later the scale difference between laser scanning
data and photogrammetric point cloud was solved by the least
squares method. The mathematical model was the 7-parametric
transformation:
The camera applied was Nikon D100 and the applied image size
was 2240 by 1488 pixels. The camera was calibrated using
calibration targets provided by the iWitness (Fraser and Hanley,
2004). One image block was acquired at each step of the load
tests.
1
X T  R ( X X S ) ,

274
(1)
IAPRS Volume XXXVI, Part 5, Dresden 25-27 September 2006
where X T is the transformed 3D point, is a scale factor, R is
a 3D rotation matrix, X S is shift and X is the original 3D point.
selecting small sample areas from two different scans. These
laser data sets were imported in the Geomagic Studio 8. The
point clouds were then transformed to surfaces. Finally, two
surfaces were registered together revealing the mean shift of
0.012 m. The final registration was, however, calculated using
common points extracted from laser data sets.
The scale was eliminated from the adjustment when comparing
data sets from different load test phases in order to prevent
deformations to adjust in the scale. 3D point clouds were not
registered to the ground truth. Therefore, difference bars in
Figures 11-14 are relatively correct, but they can include some
uniform shifts.
The inner corners of plywood triangles (Figure 4) were
observed from 3 to 8 images. In addition, the free-shaped rim
arcs from the sides were measured. Overall RMS errors (1sigma) for image measurements were 0.408 mm (initial state),
0.616 mm (the 20 kg load) and 0.575 mm (the 100 kg load). All
photogrammetric blocks were free-networks without any
reference points.
Figure 6. Moisture on the surfaces of the reference spheres
caused the distance measurements to fail. The spheres were
heavily distorted and the automatic adjustment of sphere object
failed.
4. RESULTS
Totally six 3D models were measured from the HDW pavilion:
three with photogrammetric methods and three with laser
scanning. The first models with both methods were from the
original shape of the pavilion and the next models, respectively,
were with load on the structures. In Figure 7, examples of the
models before load test are presented.
Figure 4. Detail of photogrammetric measurements.
Corresponding points, compared to 3D points from
photogrammetric block, were extracted manually from laser
scanning data (Figure 5). The corners of triangles appeared to
be too uncertain for using breakline detection algorithms and
planes were too narrow, noisy or badly oriented preventing
automatic methods for accurate corner detection. Therefore, the
small area close to a corner was selected and then the mean
point was calculated. It would have been possible to adjust 3D
points interactively at more accurate locations, but the time for
the measurements would have increased significantly. It was
assumed, if all points from laser data were chosen in the similar
way, they would be comparable. The corner points were
selected manually using the Faro Scene software.
Figure 7. Left: Photogrammetric model of the pavilion. Right:
Corresponding points extracted from laser scanning data. The
locations of two laser scanner positions are visible. Both models
are acquired before the load test.
The 3D point cloud from photogrammetric measurements was
compared with the design model. The point clouds were
registered using the least squares method. The photogrammetric
model and corresponding CAD design model were visualized in
the AutoCAD (Figure 8). The maximum difference was 0.035
m, the mean was difference 0.014 m and the standard deviation
was 0.0065 m.
Figure 5. Corresponding points to photogrammetric
measurements were selected from laser point cloud using the
Faro Scene program.
An approximate registration of two laser scans was calculated
using target spheres (Figure 6). The spheres under open sky
were covered with humidity. This caused unexpected
phenomena when the spheres were heavily distorted (Figure 6).
Only very narrow belts of the correct surfaces of the spheres
remained. The automatic fitting of spheres in the Faro Scene
software failed to be accurate because of the distorted target
spheres. Therefore, the registration was considered to be only an
approximation. The displacement of registration was tested
Figure 8. Visualization of aligned design (CAD, white) and asbuilt (photogrammetric, black) models.
275
ISPRS Commission V Symposium 'Image Engineering and Vision Metrology'
Extracted 3D points from the laser scans were compared with
the photogrammetric point clouds. After the registration, the
behaviour of the residuals acted as expected. The locations of
laser-derived corners were shifted, but very regularly according
to the corner orientation (Figure 9). The results of all
comparisons are gathered in Table 1. It must be noted that the
laser scanner represents surfaces more accurately and the shifts
were partly caused by difficulties to define accurate breaklines
and corners from laser data.
significant. Residuals on horizontal corners were smaller than
on vertical ones (Figure 11).
Figure 11. Residuals in z-direction when laser data and
photogrammetric data were compared. Corner orientations are
visible on residual pattern.
The photogrammetric observations and laser data were
compared separately before and after the load test. The results
are shown in Table 1. The residuals in z-direction are visualized
in Figure 12. The photogrammetric inspection of the 20 kg load
test case did not reveal significant deformation on the HDW
pavilion. Neither the laser scanning showed clear deformations.
The variation of differences can be explained with less accurate
interpretation with the laser scanning data.
Figure 9. Perspective views of residuals from above of the
pavilion Extracted 3D points from two different laser scans
were compared with the photogrammetric point clouds. The
residuals acted regularly according the corner orientation.
When two scans were registered together, no regular behaviour
of residual vectors according to the corner orientation could be
found anymore (figure 10). For the registration, common points
were selected from two laser data sets. The mean residual vector
length was 0.0057 m, the standard deviation was 0.0025 m and
the maximum value was 0.012 m.
Figure 12. Left: Photogrammetic data after applying the 20 kg
load on the top of the pavilion was compared with
photogrammetric initial state data. Right: Corresponding
comparison between laser data sets. Note that the coordinate
systems were different with photogrammetry and laser scanning.
Figure 10. The residual vectors of common corners are acting
randomly when two laser scan are adjusted. Residuals are
multiplied with 50 in order to ensure visibility.
Table 1. The results of all comparisons.
Laser init. –
image init.
Image init. –
Image 20 kg
Laser init. –
laser 20 kg
Image init. –
Image 100 kg
Laser init. –
laser 100 kg
Total (mm)
mean
std
max
z-direction (mm)
mean std
max
6.5
2.9
15.5
3.3
3.7
12.6
2.7
1.5
8.9
1.2
1.1
7.9
3.6
2.1
10.4
1.7
1.5
7.8
5.1
2.6
14.7
3.0
2.3
13.6
4.0
2.2
11.4
1.9
1.5
7.3
Figure 13. Left: Photogrammetric results when the 100 kg load
was placed to hang from the rim. Right: Laser scanning
comparison.
Photogrammetric and laser scanning measurements were
compared also in z-direction. As expected, corner orientations
and incident angles were visible from the residuals. On the
upper parts of the pavilion, all corners are located nearly on the
plane that is perpendicular to the z-axis. Therefore, the
interpretation error had no significant affect to the z-residuals.
However, on the vertical sides the corner orientation is
Figure 14. The photogrametric comparison of Figure 13 (left) is
recalculated with a separate scale factor in z-direction. The
276
IAPRS Volume XXXVI, Part 5, Dresden 25-27 September 2006
differences are much closer to the laser scanning case (Figure
13, right) than the original comparison (Figure 13, left) was.
5. CONCLUSIONS
The measurements revealed that the HDW pavilion meets
relatively closely the original design plans. The mean difference
was only 0.014 m and the maximum difference was 0.035 m.
When the 100 kg weights were attached to the HDW pavilion,
both methods revealed more differences than with the 20 kg
case (Table 1). The photogrammetric results (Figure 13, left)
were behaving unexpectedly compared to laser scanning results
(Figure 13, right). The shape and distribution of differences
indicated that some scale problems might have occurred
between coordinate axes, especially in z-direction. In Figure 14,
the registration was calculated leaving also the z-scale as a free
parameter resulting the mean residual vector length of 0.0042
m, the standard deviation of 0.0023 m and the maximum value
of 0.0136 m. If these values and visual plot (Figure 14) were
compared with laser scanning results (Table 1) (Figure 13,
right), the similarity was pronounced. However, leaving the zscale as a free parameter also part of the true deformation were
expected to be compensated in adjustment. Therefore, this result
was only suggestive and not verified without additional control,
such as tacheometer measurements.
The 20 kg load on the top of the HDW pavilion did not cause
such distortion that could have been verified from both
photogrammetic measurements and laser scanning data. The
longer residuals from laser data were expected, because of the
difficulties to identify corners accurately.
The 100 kg load on the rim arc of the pavilion did cause
observable variation on corresponding point differences. Both
methods revealed similar trends, however, the residuals in laser
data were not as regular as in photogrammetic data. The length
and regularity of residuals in photogrammetric data gave reason
for suspecting scale deformation within the image block.
According the both methods the side rims kept their shapes very
well. A cross-section along the rib revealed very minor
deformations.
In the Geomagic Studio 8, two surfaces can be compared in 3D.
In Figure 15, one scan before load tests and one after the 100 kg
test were compared. This examination showed, in which parts of
the HDW pavilion the deformations have occurred. One
example of the rendered surface model can be seen in Figure
16.
Both photogrammetry and laser scanning gave similar results.
In this examination, only the main node points were extracted
or measured. This method was not optimal for laser scanning
data, because it is difficult to mark a corner accurately.
Photogrammetric model after applying the 100 kg load test was
suspected to have scale distortion in z-direction. It is
recommended to use some reference targets for preventing the
possible deformations of the free image block.
The strength of laser scanning was the amount of the data
representing the detailed 3D shape of the HDW pavilion. On the
other hand, the huge amount of the data caused high
requirements for post-processing. The post-processing time with
both methods was, as expected, significantly longer than the
data acquisition.
6. REFERENCES
Alshawabkeh, Y. and N. Haala, 2004. Integration of Digital
Photogrammetry and Laser Scanning for Heritage
Documentation. XXth ISPRS Congress, Istanbul, July 2004,
IAPRS, Vol. XXXV, Part B. In: http://www.isprs.org/
istanbul2004/comm5/papers/590.pdf (visited 28.6.2006).
Figure 15. The comparison of scans before load tests and after
the 100 kg test. The 100 kg weight was attached to the left rim
of the pavilion. Note: Result is not comparable to other
difference images on this paper. The pavilion is now looked
from below, unlike in the other difference examinations in this
paper.
Cardenal J., E. Mataa M. Ramos, J. Delgadoa, M.A.
Hernandeza, J.L. Pereza, P. Castroa, and M. Torresa, 2005.
Low cost digital photogrammetric techniques for 3D
modelization in restoration works. A case study: St. Domingo
deSi
l
os
’c
hur
c
h(
XI
Vt
hc
e
nt
ur
y
,Al
c
a
l
al
aRe
a
l
,Spa
i
n)
,CI
PA
2005 XX International Symposium, 26 September – 01
October, 2005, Torino, Italy. In: http://cipa.icomos.org/
fileadmin/papers/Torino2005/722.pdf (visited 27.6.2006).
El-Hakim, S., E. Whiting, L. Gonzo, and S. Girardi. S. 2005.
3D Reconstruction of Complex Architectures from Multiple
Data. Proceedings of the ISPRS Working Group V/4 Workshop
3D-ARCH 2005: "Virtual Reconstruction and Visualization of
Complex Architectures" Mestre-Venice, Italy, 22-24 August,
2005. Volume XXXVI, Part 5/W17 ISSN 1682-1777. In:
http://www.commission5.isprs.org/3darch05/pdf/1.pdf (visited
28.6.2006).
Figure 16. Rendered surface model from a single laser scan. The
glass parts were eliminated, because laser scanning could not
record them accurately.
277
ISPRS Commission V Symposium 'Image Engineering and Vision Metrology'
Fraser, C.S and H.B. Hanley, 2004. Developments in closerange photogrammetry for 3D modelling: the iWitness example,
International Workshop: Processing & Visualization using
High-Resolution Imagery, Pitsanulok, Thailand, 18-20 Nov. In:
http://www.photogrammetry.ethz.ch/pitsanulok_workshop/pape
rs/09.pdf (visited 27.6.2006).
load tests in civil engineering material testing. In: 6th
Conference on Optical 3-D Measurement Techniques. Prof. A.
Grün, Institute of Geodesy and Photogrammetry, Swiss Federal
Institute of Technology, Zürich, 22-25.9.2003, Vol. II, pp. 8088.
Pollefeys, M., M. Vergauwen and L. Van Gool, 2000.
Automatic 3D modeling from image sequences, invited
presentation, IAPRS, Vol. XXXIII, Part B5, pp. 619-626.
Fröhlich, C. and M. Mettenleiter, 2004. Terrestrial LaserScanning - New Perspectives in 3D-Surveying. Proceedings of
the ISPRS working group VIII/2 'Laser-Scanners for Forest and
Landscape Assessment' Freiburg, Germany 03-06 October
2004. ISSN 1682-1750 VOLUME XXXVI, PART 8/W2. In:
http://www.isprs.org/commission8/workshop_laser_forest/FRO
EHLICH.pdf (visited 27.6.2006).
Schindler, K. and J. Bauer, 2003. A Model-Based Method For
Building Reconstruction. IEEE 1st International Workshop on
Higher-Level Knowledge in 3D Modeling and Motion Analysis
(HLK 2003), 17 October 2003, Nice, France, Proceedings.
IEEE Computer Society 2003, ISBN 0-7695-2049-9. In:
http://lear.inrialpes.fr/people/triggs/events/iccv03/cdrom/hlk03/
schindler.pdf (visited 27.6.2006).
Gordon, S., D. Lichti, J. Franke, and M. Stewart, 2004.
Measurement of Structural Deformation using Terrestrial Laser
Scanners. 1st FIG International Symposium on Engineering
Surveys for Construction Works and Structural Engineering
Nottingham, United Kingdom, 28 June –1 July 2004. In:
http://www.fig.net/nottingham/proc/ts_03_2_gordon_etal.pdf
(visited 27.6.2006).
Schuhmacher S. and J. Böhm, 2005. Georeferencing of
terrestrial laserscanner data for applications in architectural
modelling. Proceedings of the ISPRS Working Group V/4
Workshop 3D-ARCH 2005: "Virtual Reconstruction and
Visualization of Complex Architectures" Mestre-Venice, Italy,
22-24 August, 2005. Volume XXXVI, Part 5/W17
ISSN 1682-1777. In: http://www.commission5.isprs.org/
3darch05/pdf/15.pdf (visited 27.6.2006).
Gruen, A., F. Remondino and L. Zhang, 2002. Reconstruction
of the Great Buddha of Bamiyan, Afghanistan. In: International
Archives of Photogrammetry and Remote Sensing, Vol.
XXXIV, part 5, Corfu, Greece, pp. 363-368.
Sternberg, H,. Th. Kersten, I. Jahn, and R. Kinzel, R. 2004.
Terrestrial 3D laser scanning –data acquisition and object
modelling for industrial as-built documentation and
architectural applications. XXth ISPRS Congress, Istanbul, July
2004, IAPRS, Vol. XXXV, Commission VII, Part B2, pp. 942947.
Grussenmeyer, P. and Yasmine, J., 2003. The Restoration of
Beaufort Castle (South-Lebanon): A 3D Restitution According
to Historical Documentation. In XIXth CIPA International
Symposium, Antalya, Turquey. Sept. 30th. Oct 4th, 2003.
IAPRS, Vol. XXXIV-5/C15 ISSN 1682-1750 pp. 322-327.
Guarnieri, A., F. Pirotti, M. Pontin. and A. Vettore, 2005.
Combined 3D surveying techniques for structural analysis
applications. Proceedings of the ISPRS Working Group V/4
Workshop 3D-ARCH 2005: "Virtual Reconstruction and
Visualization of Complex Architectures" Mestre-Venice, Italy,
22-24 August, 2005. Volume XXXVI, Part 5/W17
ISSN 1682-1777. In: http://www.commission5.isprs.org/
3darch05/pdf/28.pdf (visited 27.6.2006).
Thies, M. and H. Spiecker, 2004. Evaluation and future
prospects of terrestrial laser scanning for standardized forest
inventories. IAPRS, Vol. XXXVI, part 8/W2, pp. 192-197.
Valiev, I. 1999. 3D Reconstruction of Architectural
Objects from Photos. The 9-th International Conference
on Computer Graphics and Vision, Moscow, Russia,
August 26 - September 1, 1999 In: http://www.integra.jp/
eng/papers/reno_art.pdf (visited 28.6.2006).
S. N. Harun and A. G. Ahmad, 2003, The Restoration of
Suffolk House, Penang, Malaysia, CIPA 2003 XIXth
International Symposium, 30 September –04 October, 2003,
Antalya, Turkey. In: http://cipa.icomos.org/fileadmin/papers/
antalya/25.pdf (visited 27.6.2006).
Werner, T., F. Schaffalizky and A. Zisserman, 2001. Automated
Architecture Reconstruction from Close-range Photogrammetry.
Proc. on CIPA 2001 International Symposium: Surveying and
Documentation of Historic Buildings -- Monuments -- Sites,
Traditional and Modern Methods (2001), p. 54-63.
Lee I. and Y. Choi, 2004. Fusion of Terrestrial Laser Scanner
Data and Images for Building Reconstruction. XXth ISPRS
Congress, Istanbul, July 2004, IAPRS, Vol. XXXV,
Commission V. In: http://www.isprs.org/istanbul2004/comm5/
papers/189.pdf (visited 27.6.2006).
7. ACKNOWLEDGEMENTS
The financial support of the Academy of Finland (projects
121053 and 211476) and Ministry of the Environment for the
project “
the use of ICT 3D measurement techniques for highqualityc
ons
t
r
uc
t
i
on”is gratefully acknowledged.
Lehto, A. and T. Seppänen, 2005. HDW Info Pavilion, PuuWood-Holz-Bois, No. 4, p. 26.
Lindenbergh, R., N. Pfeifer and T. Rabbani, 2005. Accuracy
analysis of the leica HDS3000 and feasibility of tunnel
deformation monitoring. ISPRS WG III/3, III/4, V/3 Workshop
"Laser scanning 2005", Enschede, the Netherlands, September
12-14, 2005. pp. 24-29.
Hampel, U. and H.-G. Maas, 2003. Application of digital
photogrammetry for measuring deformation and cracks during
278